Analysis of penicillamine using Cu-modified graphene quantum dots synthesized from uric acid as single precursor.

A simple methodology was developed to quantify penicillamine (PA) in pharmaceutical samples, using the selective interaction of the drug with Cu-modified graphene quantum dots (Cu-GQDs). The proposed strategy combines the advantages of carbon dots (over other nanoparticles) with the high affinity of PA for the proposed Cu-GQDs, resulting in a significant and selective quenching effect. Under the optimum conditions for the interaction, a linear response (in the 0.10-7.50 µmol/L PA concentration range) was observed. The highly fluorescent GQDs used were synthesized using uric acid as single precursor and then characterized by high resolution transmission electron microscopy, Raman spectroscopy, X-ray diffraction, Fourier transform infrared spectroscopy, fluorescence, and absorption spectroscopy. The proposed methodology could also be extended to other compounds, further expanding the applicability of GQDs.

Introduction

The development and application of nanomaterials is one of the current trends of nanoscience and nanotechnology [1], [2], [3], [4]. Among those, photoluminescent nanomaterials have drawn increasing attention because they have the potential to impact a wide number of fields ranging from optoelectronics to bio-analysis [5], [6]. While the synthesis and applications of semiconductor qpan class="Chemical">uantum dots (S-QDs) [2], [7], [8], [9], [10], [11] have gained abundant recognition [11], their class="Chemical">pan class="Disease">toxicity can severely impair their use, especially for biomedical applications. Addressing this limitation, carbon-based QDs (CQDs) present themselves as an attractive alternative. Similar to traditional QDs, CQDs exhibit unique optical and electronic characteristics, including large two-photon excitation cross-sections, lack of blinking, high water solubility, biocompatibility, and low toxicity [12], [13], [14], [15]. Because these nanomaterials can be synthesized in the 2–20 nm range, they are also intrinsically photoluminescent, a property largely associated to the quantum confinement, their edge structure, and the presence of surface defects [16], [17], [18], [19], [20], [21], [22]. Different approaches have been proposed to obtain CQDs, including electrochemical synthesis [23], arc discharge [24], and laser ablation [25]. Most of these top-down approaches involve a non-selective exfoliation process that requires the use of toxic reagents and/or specific facilities. Bottom-up approaches involving the carbonization of various organic precursors (citric acid, glucose, sucrose, glycol, ascorbic acid, etc.) have also been applied to obtain luminescent carbon-based nanomaterials [5], [26], [27]. However, these approaches typically require several steps to control their water solubility and luminescence properties [25]. An additional limitation affecting both methods is that the resulting particles often contain a significant proportion of amorphous carbon, thus limiting their use. Aiming to address these drawbacks, this study develops a simple, fast and economical approach to producing high quality and photoluminescent graphene quantum dots (GQDs). The GQDs were synthesized via a hydrothermal process using uric acid as a single, low-cost carbon and nitrogen source. To the best of our knowledge, this is the first time that GQDs are synthesized from uric acid as single precursor (with constant C/N ratio). This methodology yields hydrophilic GQDs with competitive photoluminescent properties, and narrow size distribution which can be used for the development of a variety of fluorescent sensing analytical systems.

Among the potential targets with pharmaceutical relevance, pan class="Chemical">penicillamine (PA) is a natural compound containing class="Chemical">pan class="Chemical">sulphur amino acids derived from hydrolytic degradation of penicillin-based antibiotics. Its importance relies on its use for the treatment of hepatolenticular degeneration (Wilson´s disease) as well as other disorders (rheumatoid arthritis, primary biliary cirrhosis, scleroderma, fibrotic lung diseases, cystinuria, heavy element poisoning, and progressive systemic sclerosis) [28], [29]. Due to poor tolerance (abdominal pain, nausea, and diarrhea) and a range of adverse effects, content of PA in pharmaceutical preparations should be closely monitored. Therefore, the development of fast, simple, sensitive, and selective methods for the determination of this analyte is very important from not only the therapeutic but also the regulatory standpoint. Chemically, PA is considered as a strong chelating agent and presents a good reactivity with most heavy metals and forms stable complexes with copper ions [30], [31]. Taking this property into account, the present report presents a simple approach to modifying GQDs with copper (II) ions (Cu-GQDs) and using them as a simple fluorescence detection strategy, based on the synergistic quenching effect of Cu-GQDs by PA. The herein proposed strategy also aims at complementing previous reports focused on the use of other nanoparticles towards the analysis of PA [32], [33], [34], [35], [36], [37]. The feasibility of the proposed fluorescent sensor for the quantification of PA in pharmaceutical samples was also demonstrated.

Materials and methods

Chemicals and reagents

n class="Chemical">Uric acid (class="Chemical">pan class="Chemical">UA, ≥ 99.0%) was purchased from Alfa Aesar (Ward Hill, MA, USA), and sulfuric acid (ACS/FCC, BDH ARISTAR, 95.0%–98%) was acquired from VWR (Pittsburgh, PA, USA). Sodium hydroxide (98%), copper (II) chloride (99.999%), sodium acetate (≥ 99.0%), acetic acid (≥ 99.7%), and D-penicillamine (PA, ≥ 98%) were obtained from Sigma Aldrich (Germany). All aqueous solutions were prepared using 18.2 MΩ cm water, purified with a Milli-Q water system (Millipore, Molshem, France). Solvents such as methanol (99.9%), absolute ethanol (99.8%), acetone (≥ 99.9%), and acetonitrile were obtained from Panreac (Barcelona, Spain) and chloroform (≥ 99%) was obtained from Sigma Aldrich (Steinheim, Germany). All the reagents were used as received. Acetate buffer solution was prepared by mixing 0.1 mol/L sodium acetate and 0.1 mol/L acetic acid solutions in the proper proportions to get the required pH in the 3.0–6.0 range [38]. CuCl2 modified GQDs (Cu-GQDs) solution was prepared by mixing an appropriate concentration of CuCl2·H2O solution (0.1 mol/L) and GQDs solution in a 1:19 ratio (m/m). Then the mixture was stirred for 5 min and stored at room temperature in the darkness until use.

Other compounds used in the interference study were n class="Chemical">sucrose, class="Chemical">pan class="Chemical">lactose, glucose, sodium dodecyl sulphate (SDS), starch, sodium chloride, and potassium chloride; all of them were obtained from Sigma (Madrid, Spain).

Instrumentation

The morphology of the GQDs was determined using a high-resolution transmission electron microscopy (HRTEM) at an accelerating voltage of 200 kV (JEOL 2010-F Field Emission Transmission Electron Microscope). All UV–Vis absorption spectra were recorded using a spectrophotometer (SECOMAM UVI Light XS 2) equipped with a LabPower V3 50 for absorbance data acquisition using 10 mm qpan class="Chemical">uartz cuvettes. Fluorescence (FL) spectra were obtained on a fluorescence spectrophotometer (Photon Technology International (PTI) Inc. Qclass="Chemical">pan class="Chemical">uantaMaster 40) equipped with a 75-W continuous xenon arc lamp. An ASOC-10 USB interface FeliXGX software was used for data acquisition. The FL of GQDs samples was observed using a handheld UV lamp (model UVLS-28, 115 V ~ 60 Hz, 0.16 Amp.) with long (365 nm) and short (254 nm) wavelength assembly. FTIR spectra were recorded on an FTIR spectrophotometer (IRAffinity-1, Shimadzu Corporation International Marketing Division, Japan). The pH measurements were achieved in a Crison Basic 20 pH-meter with a combined glass electrode (Barcelona, Spain). The GQDs centrifugation was performed with an ultracentrifuge (Centrofriger BL-II model 7001669, J.P Selecta). X-ray diffraction (XRD) analyses were carried out on an X-ray diffraction spectrometer (XRD Rigaku Ultima IV, Rigaku Americas Corporation, USA) with parallel beam and focused beam, small angle and thin films attachment. Raman spectra were recorded in a Raman spectrometer (BWS415-785S with a probe BAC100-785, BWTek, Inc., USA) using a laser power with an excitation port of 785 nm.

Synthesis and purification of GQDs

GQDs were synthesized from n class="Chemical">UA by acidic chemical treatment. In this case, 0.5 g of class="Chemical">pan class="Chemical">UA was combined with 1 mL of concentrated sulfuric acid and allowed to react under continuous stirring at 200 °C for 1 h. The brown reaction product was allowed to cool down to room temperature and then neutralized with NaOH (to pH ~ 7) under vigorous stirring. The resulting suspension was treated with a mixture of methanol and acetone (1:4, v/v) and filtered through a 0.45 µm membrane to separate large particles. The subsequent purification steps involved the extraction of GQDs in methanol and acetone (1:4, v/v) and centrifugation at 9000 rpm (to precipitate larger particles and salts). Finally, the yellow product was dried under a N2 stream and stored at room temperature in the darkness until use.

Quantum yield (QY) measurements

n class="Chemical">Fluorescein (dissolved in 0.1 mol/L class="Chemical">pan class="Chemical">NaOH, FL QY=92.5%) was chosen as standard. The fluorescence and UV–Vis spectra of fluorescein and obtained GQDs were recorded at different concentrations, respectively. Then FL QY of GQDs in water was calculated according to the following equation (Eq. (1)) [39], [40]:where QYS and QYSt represent the quantum yield of the sample and the standard, respectively, K is the slope determined by the fluorescence intensity versus the UV–Vis intensity curves, and η is the refractive index of the solvent. In order to minimize the re-absorption effects, absorption in the 10 mm cuvette was kept below 0.10 at the excitation wavelength.

Sample preparation and analytical procedure

For the preparation of pharmaceutical sn class="Chemical">amples, 3 capsules of Cupripén® 250 (class="Chemical">pan class="Chemical">D-Penicillamine, Laboratorios Rubió, S.A., Barcelona, Spain) were accurately weighed and mixed. A portion of the powder obtained, equivalent to the average weight of a capsule, was dissolved in 500 mL of deionized water. Then, the obtained suspension was filtered through a 0.45 µm filter and the filtrate was used to prepare the working solution of PA for the analysis. The interaction studies were performed by recording the fluorescence spectra of Cu-GQDs, 1:19 ratio, in the absence and in the presence of different concentrations of PA in acetate buffer (5 mmol/L, pH 6.0), after 5 min of reaction at room temperature, and preserved from the light. The fluorescence spectrum was measured at excitation wavelength of 362 nm and collecting the emission between 380 and 640 nm. Both the excitation and the emission slit widths were 2 nm. I0/I was used as analytical signal, where I0 and I were the fluorescence intensity at 452 nm of the systems in the absence and in the presence of PA, respectively. When samples were analyzed, PA standard solution was replaced by the pharmaceutical samples prepared as described at the beginning of this section.

Results and discussion

Characterization of GQDs

The GQDs herein reported were synthesized using n class="Chemical">UA as single precursor and different characterization techniques were used in order to assess the properties of the GQDs obtained. Optical properties in aqueous media were studied through UV–Vis and FL spectra. As can be observed in Fig. 1A, an absorption peak was observed at 260 nm, typically attributed to π-π* transitions of the aromatic C=C class="Chemical">pan class="Chemical">sp2 domains [41], [42]. The broader band (observed at approximately 380 nm) may be attributed to n-π* transitions of C=O bond [43] and the trapping of excited-stage energy by the surface states. The fluorescence spectrum shows a strong emission peak at 452 nm when excited at 362 nm, and the full width at half maximum (FWHM) is about 65 nm, which is similar to the value previously reported for other CQDs [6], [16], [41], [44], [45], [46]. In addition, the fluorescence spectrum at different excitation wavelengths shows a maximum intensity at 364 nm. Inserts in Fig. 1A include photographs of the water-dispersed GQDs under either visible or UV light (365 nm). A diluted aqueous GQDs solution shows a faint yellow color, and an intense blue fluorescence color under UV light, which can be easily observed by the naked eye. To further characterize the optical properties of the prepared GQDs, the FL spectra of GQDs were recorded at different excitation wavelengths (Fig. 1B), showing the maximum emission fluorescence at an excitation wavelength of 364 nm. In addition, a slight excitation-dependent fluorescence behavior was observed, shifting to longer wavelengths without a significant change in shape. This fact suggests the presence of abundant surface defects, approximating the trend reported for most of the reported carbon-based nanomaterials. These results also imply that discrete energy levels are present in these nanomaterials [16], [41], [44], [45]. The blue emission was attributed to the free zig-zag effect with a carbine-like triplet ground state of σ1π1 [47], [48] and their blue luminescence was generated from intrinsic states in the highly crystalline structure. In addition, the effect of increasing concentration of GQDs in aqueous solution on the photoluminiscence intensity was also evaluated. The results showed a linear relationship between photoluminescence intensity and the concentration of nanoparticles in the 0–0.1 mg/mL range.

Fig. 1

(A) UV–Vis spectrum (red line), excitation (blue line) and fluorescence spectra (black line) of obtained GQDs, inset: photographs of GQDs solution under visible (left) and 365 nm UV (right) light; (B) fluorescence spectra of GQDs at 300–400 nm excitation wavelengths.

The absorbance and FL intensity of GQDs at different pH values was also investigated. It was observed that the FL emission intensity of obtained GQDs was pH-dependent, as it was previously reported for this kind of nanoparticles [41], [49]. Significant changes in both the FL intensity and the maximum emission wavelength at different pH values were observed (Fig. S1). The FL intensity was increased from pH 2 to pH 6, reaching a maximum value at pH 6. At higher pH values (in the 6–10 range), a blue shift from 452 to 445 nm with decreased FL intensity was observed. This fact can be attributed to the deprotonation of carboxyl groups of G-QDs in alkaline solution, increasing the electrostatic repulsions between them, and overcoming the trend of aggregation through layer-layer stacking [50], [51]. This pH-dependent behavior suggests that the origin of photoluminiscence can be linked to small pan class="Chemical">sp2 fragments conjugated with class="Chemical">pan class="Chemical">oxygen-containing functional groups such as carboxylic groups in the GQDs, where the charge and corresponding electronic distribution depend on the pH of the solution [52].

The solubility and fluorescence intensity of GQDs in different solvents was also determined. To this end, powder GQDs was added to several solvents, such as deionized pan class="Chemical">water, class="Chemical">pan class="Chemical">glycerol, methanol, ethanol, acetone, acetonitrile, and chloroform. The photographs of GQDs in the several solvents, under visible and 365 nm UV light, demonstrated that solutions of these nanoparticles were stable in deionize water and did not show evidence of sedimentation or agglomeration. While these GQDs can also disperse in other polar organic solvents besides the deionized water, such as glycerol, ethanol, and methanol (Fig. S2), they displayed limited solubility in acetonitrile and acetone and were insoluble in chloroform. Similar to the pH effect, this behavior can be related to the nature of the functional groups of the GQDs. In this context, carbonyl groups in acetone are typical Lewis acid whereas hydroxyl groups in methanol, ethanol, and glycerol are a typical Lewis base [26]. This fact enables to use the obtained GQDs for non-aqueous applications in different fields.

The morphology of prepared GQDs was investigated using HRTEM images, and representative results are shown in Fig. 2A. It can be clearly observed that the produced GQDs consist of small individpan class="Chemical">uaclass="Chemical">pan class="Gene">l particles with a narrow size distribution, ranging between 2.5 and 4.8 nm; with an average value of 3.7 nm (Fig. 2B). This finding suggests that the GQDs consist of dots composed of graphene layers (Fig. 2C), uniformly arranged without apparent aggregation. The HRTEM images also show a lattice fringe of 0.23 nm, in good agreement with the crystal structure of graphene (consistent with the basal plane distance of bulk graphite and the hexagonal lattice) [53]. Moreover, due to their small size, the edge structures such as zig-zag and armchair edges, are more evident based on their HRTEM images, confirming the hexagonal honeycomb structures and bond lengths (1.84 Å of C–C) of graphene (Fig. 2D). These special edges and shapes of the GQDs are responsible for their optical and electronic properties.

Fig. 2

HRTEM image of (A) GQDs, (B) GQDs size distribution, (C) morphology, and (D) lattice edges (schematic illustration showing the orientation of the graphene and zig-zag edge of a dot).

HRTEM image of (A) GQDs, (B) GQDs size distribution, (C) morphology, and (D) lattice edges (schematic illustration showing the orientation of the n class="Chemical">graphene and class="Chemical">pan class="Chemical">zig-zag edge of a dot).

The surface and chemical composition of prepared GQDs were determined by FTIR spectra. Fig. 3A shows the FTIR spectra of n class="Chemical">UA (precursor) and the obtained GQDs. The results reveal the presence of characteristic functional groups in both the precursor and the GQDs. In this way, it was possible to detect the stretching vibration bands of O-H at 3401 cm−1, which indicates the presence of hydroxyl groups and the vibrational absorption band of C–O groups at 1063 cm−1. The presence of the stretching vibration modes of C–H at 3031 cm−1 and 1433 cm−1, the vibrational absorption band of C=O conjugated with condensed aromatic class="Chemical">pan class="Chemical">carbons at 1671 cm−1, the vibrational absorption band of C–O at 1063 cm−1 , and the weak vibration absorption band of N–H at 1595 cm−1 were also observed [22], [40], [54]. It is important to point out that these nitrogen-containing functional groups were considered essential to support the interaction with the copper ions, which would be otherwise released from the surface of the GQDs upon the addition of PA. The XRD pattern in Fig. 3B shows a series of crystalline peaks, forming a broad peak centered at around 29.4° and an interlayer spacing (calculated) of 0.304 nm. This value is broader than that of graphite [55] and it is attributed to the ordered stacking of obtained GQDs.

Fig. 3

(A) FTIR spectra of uric acid (red line) and obtained GQDs (black line); (B) XRD patterns of powder GQDs; and (C) Raman spectra of uric acid (red line) and obtained GQDs (black line).

(A) FTIR spectra of pan class="Chemical">uric acid (red line) and obtained GQDs (black line); (B) XRD class="Chemical">patterns of powder GQDs; and (C) Raman spectra of class="Chemical">pan class="Chemical">uric acid (red line) and obtained GQDs (black line).

Raman spectroscopy was used to investigate the structure of the GQDs prepared from n class="Chemical">UA. Representative results are shown in Fig. 3C, where two characteristic peaks at 1385 cm−1 (D band) and 1580 cm−1 (G band) were expected. As it can be observed, the D band, resulting from the vibrations of class="Chemical">pan class="Chemical">carbon atoms with dangling bonds in the termination plane of carbon structures, clearly dominates the spectrum. It is important to point out that the G band, typically associated with the E2g vibrational mode of aromatic domains in the two-dimensional hexagonal lattice [56], was barely noticeable, an effect that has been associated with the presence of nitrogen atoms [57]. Therefore, considering that graphitic structures were systematically observed by microscopy (Fig. 2A), the ratio of the intensities was used to estimate relative abundance of carbon structures [58]. In our case, a ID/IG ratio of 1.35 was calculated for the obtained GQDs, indicating the presence of abundant defects on the obtained GQDs with a partially disordered nanocrystal structure, arising from the small sp2 cluster size [44].

To further demonstrate the potential advantages of generating GQDs from pan class="Chemical">UA, the qclass="Chemical">pan class="Chemical">uantum yield of these nanoparticles was also calculated, according to the procedure detailed in Section 2.4. In this case, the FL QY of the GQDs resulted to be 44.4%. This value is about 30% higher than that of most conventional GQDs reported in bibliography (Table S1). This enhancement of QY can be attributed to the density of carbene structures in the zig-zag edges, showing interesting luminescence with high quantum yield. Furthermore, compared to other reported GQDs syntheses which involve the use of several precursors, such as citric acid and urea (Table S1), the proposed methodology avoids the amidation reaction due to the use of a single precursor containing a constant (and appropriate) carbon/nitrogen ratio. Therefore, these facts suggest that the optical properties of GQDs are influenced by a combination of factors including their size, shape, and functionalization.

Effect of PA on the photoluminescence response of GQDs

Preliminary interaction studies between the GQDs obtained from n class="Chemical">UA and PA were carried out by recording the FL spectra of GQDs in the absence and in the presence of PA. It was observed that the presence of PA did not change the FL intensity of GQDs. Thus, it was assumed that PA did not interact with GQDs. However, interesting interactions between synthesized GQDs, previously modified with class="Chemical">pan class="Chemical">copper (II) ions (Cu-GQDs), and PA were observed (Fig. S3). It is known that PA is a strong chelating agent presenting good reactivity with copper ions. Thus, the interaction between Cu-GQDs and PA can be studied based on the changes in the FL intensity of modified GQDs. In this regard, the optimization of several chemical variables on the reaction between Cu-GQDs and PA was carried out to obtain the most favorable conditions for the interaction with PA.

In order to evaln class="Chemical">uate the effect of class="Chemical">pan class="Chemical">copper (II) ions on the FL intensity of synthesized GQDs, different concentrations of copper (II) ions between 0.1 and 5.0 mmol/L (13.5–672.3 mg/L) were added to the aqueous solution of GQDs. Then, the FL intensity of the Cu-GQDs system was recorded. It was observed that the FL decreased to about 30% of its initial value when the concentration of copper (II) ions added to GQDs solution was 0.50 mmol/L (Fig. 4A). Higher copper (II) concentration did not produce additional quenching. Therefore, in order to avoid the use of high concentrations of cupric ions, the effect of CuCl2/GQDs ratio in the PA sensing was also investigated. For this purpose, different CuCl2/GQDs ratios in the 1:3 to 1:47 range in the absence (I0) and in the presence (I) of 10 µmol/L PA were studied (Fig. 4B). A maximum analytical signal was found when the CuCl2/GQDs ratio of 1:19 was used. Therefore, this ratio was selected as optimum and was used for the determination of PA in pharmaceutical preparations.

Fig. 4

(A) Effect of increasing concentrations of CuCl2 on the fluorescence intensity of GQDs, and (B) effect of CuCl2/GQDs ratio (CPA=10 µmol/L) in the Cu-GQDs-PA interaction.

(A) Effect of increasing concentrations of n class="Chemical">CuCl2 on the fluorescence intensity of GQDs, and (B) effect of class="Chemical">pan class="Chemical">CuCl2/GQDs ratio (CPA=10 µmol/L) in the Cu-GQDs-PA interaction.

The selection of the buffer composition was a crucial variable for the reaction. Preliminary studies were conducted using different buffer compositions such as pan class="Chemical">citrate, class="Chemical">pan class="Chemical">citrate/phosphate and acetate/acetic acid buffer. A destabilization of the reaction in the presence of citrate and citrate/phosphate buffer was observed, probably due to the reaction with the analyte. However, when acetate/acetic acid buffer was used a stabilization of the system was observed. Therefore, this buffer was used for all subsequent experiments. Moreover, the effect of pH on the FL intensity of the Cu-GQDs system was also evaluated using acetate/acetic buffer ranging from pH 3.0 to 6.0 in the absence (I0) and in the presence of 10 µmol/L PA (I). I0/I was used as the analytical signal. It was found (data not shown) that the pH significantly influenced the FL intensity of the Cu-GQDs-PA system, resulting in a progressive enhancement of I0/I when the pH was increased and reaching a maximum I0/I value at pH 6. Therefore, this value was selected as optimum for the subsequent experiments. Additionally, the effect of the buffer concentration of 2–10 mmol/L at pH 6 in Cu-GQDs-PA system was also investigated. No significant differences were found in the I0/I. However, slight differences in the stabilization time of the system were observed. Specifically, a faster stabilization of the system was observed when the buffer concentration was 5 mmol/L. Higher buffer concentrations led to slower stabilization of the system. Therefore, 5 mmol/L was selected as optimum buffer concentration. On the other hand, studies on the reaction time showed (data not shown) that FL intensity remained constant after stirring the mixture for 30 s and performing the measurements after 5 min of reaction.

FL detection is a useful alternative for the sensitive monitoring of the concentration of PA. Therefore, the analytical use of pan class="Chemical">carbon-based nanomaterials, such as GQDs, can play an important role in the detection of compounds of interest. This fact is due to the presence of class="Chemical">pan class="Chemical">sp2 bonded carbon lattices and surface functional groups, which facilitate the interaction of other moieties to their surface through electrostatic interaction, π–π stacking, and chemical reactions. It is known that metallic ions (such as Hg2+, Fe3+ and Cu2+) can bind with carbon-based QDs through surface functional groups or even through surface adsorption quenching the fluorescence of GQDs [59], [60], [61]. Moreover, PA can coordinate with metal ions (such as Cu2+) through complexation reaction. Thus, metallic ions may act as a linking agent between GQDs and PA molecules to bring them into proximity of each other, leading to the fluorescence quenching of the Cu-GQDs system. In this context, the Cu-GQDs system can be used as fluorescent probe for the determination of PA.

Analytical features for the determination of PA

In order to investigate the possible analytical applicability of the obtained GQDs in the presence of pan class="Chemical">copper (II) ions, as fluorescent sensing systems, several analytical performance characteristics were evalclass="Chemical">pan class="Chemical">uated. PA was used as an analyte under the optimized experimental conditions and the procedure was described in Section 2.5. As it can be observed, with the increase of PA concentration, a progressive decrease in the fluorescence of Cu-GQDs was observed (Fig. 5).

Fig. 5

Fluorescence quenching of Cu-GQDs in the absence and in the presence of different concentrations of PA in the 0.1–20 µmol/L range. Insert: the linear relationship between the relative fluorescence intensity and PA concentration.

I0/I was used as a possible analytical signal, where I0 and I were the fluorescence intensity at 452 nm of the systems (exciting at 364 nm) in the absence and in the presence of PA, respectively. The obtained experimental data were satisfactorily fitted according to a Stern-Volmer plot eqpan class="Chemical">uation (Eq. (2)), obtaining a good relationship (r>0.997) in the 0.10–7.5 µmol/L range:

The precision of the methodology was evaln class="Chemical">uated in terms of repeatability and reproducibility. To determine the repeatability of the methodology, 10 analyses of sclass="Chemical">pan class="Chemical">amples containing 1 µmol/L of PA were carried out and the obtained relative standard deviation (R.S.D.) was 1.51%. Then, the reproducibility was estimated for three replicates of 1 µmol/L PA under inter-day conditions (for two consecutive days), obtaining an R.S.D. of 8.02%. A limit of detection (LOD) of 0.03 µmol/L was obtained for PA determination, based on the IUPAC method (blank signal plus 3 times its standard deviation, n=10)

In order to apply the proposed analytical method to pharmaceutical formulations, a selectivity study was performed considering the uspan class="Chemical">ual excipients present in these types of sclass="Chemical">pan class="Chemical">amples. While many studies have reported that the fluorescence of QDs can be affected by metallic cations, it is unlikely that pharmaceutical samples could be contaminated with metallic cations. For this reason, the interference study was performed by adding the possible interfering compounds to a solution containing Cu-GQDs and 1 µmol/L of PA. Possible interfering compounds of pharmaceutical formulations such as magnesium stearate, quinolone, and titanium oxide were eliminated by introducing a previous filtration step of the sample since these compounds are water-insoluble. Moreover, other common excipients such as sucrose, fructose, lactose, glucose, maltose, starch, SDS, K+, and Na+ were studied. The results showed that compounds typically present in these samples (in a 1:5 concentration ratio) only provided marginal changes (<4%) in the analytical signal of Cu-GQDs (Fig. 6). While it is reasonable to expect cross-reactivity with other copper-quelating drugs like tetramines (also used in Wilson's disease [62]), or compounds used in general chelation therapy [63], we do not expect pharmaceutical preparations containing PA to be contaminated with any of these compounds.

Fig. 6

Relative fluorescence intensity (I/I0) of Cu-GQDs in the absence and in the presence of 1 µmol/L of PA and 5 µmol/L of different coexisting compounds (sucrose, fructose, glucose, maltose, SDS, starch, lactose, Na+ and K+).

Relative fluorescence intensity (I/I0) of Cu-GQDs in the absence and in the presence of 1 µmol/L of PA and 5 µmol/L of different coexisting compounds (n class="Chemical">sucrose, class="Chemical">pan class="Chemical">fructose, glucose, maltose, SDS, starch, lactose, Na+ and K+).

Analytical application

In order to demonstrate the potential applicability of the proposed sensing system for pharmaceutical analysis, the determination of PA in a pharmaceutical sn class="Chemical">ample (class="Chemical">pan class="Chemical">Cupripen® 250) was performed. This sample was analyzed by triplicate according to the procedure described in Section 2.5. To evaluate the matrix effect, the standard addition method was also used for the determination of PA in the sample and recoveries were examined. The PA content in the real sample was obtained from the standard curve and the regression equation. The results are summarized in Table 1.

Table 1

Determination of PA in pharmaceutical formulation (Cupripen® 250 mg, Laboratorios Rubió, S.A., Barcelona, Spain) by standard addition method (n=3). In all cases, the amount of PA in sample declared by manufacturer was diluted to 0.335 µmol/L.

Amount of standard PA added (µmol/L)	Total PA found (µmol/L)	% Recovery	R.S.D. (%)
0.500	0.88±0.01	112.6	0.8
1.500	1.87±0.02	110.3	1.0
2.000	2.31±0.03	92.3	1.4

Determination of PA in pharmaceutical formulation (pan class="Chemical">Cupripen® 250 mg, Laboratorios Rubió, S.A., Barcelona, Sclass="Chemical">pain) by standard addition method (n=3). In all cases, the amount of PA in sclass="Chemical">pan class="Chemical">ample declared by manufacturer was diluted to 0.335 µmol/L.

As it can be observed, recoveries of PA in the pharmaceutical span class="Chemical">ample were between 92.3% and 112.6%. The results obtained by the proposed analytical method were comclass="Chemical">pared with the value declared by the manufacturer using the Student statistical t-test. No significant differences were observed.

As such, the developed sensor provides a simple avenue for the determination of PA in different pharmaceutical formulations without the interference of common excipients present in the span class="Chemical">ample. The analytical performance of the proposed methodology presents clear advantages in terms of simplicity and cost over traditional analytical approaches (like HPLC) or other fluorescent probes based on the interaction of PA with class="Chemical">pan class="Chemical">copper ions [64], [65]. In addition, the proposed strategy also contributes to complementing previous reports focused on the use of nanomaterials towards the analysis of PA (Table S2).

Conclusions

A simple, fast, low-cost, high-yield, and acidic-based treatment method for the synthesis of blue photoluminescence GQDs using a single organic material was developed and the analytical applicability of GQDs was demonstrated. The proposed methodology is suitable for obtaining GQDs with a narrow size distribution (2.5–5.0 nm) which exhibit a characteristic blue luminescence and a competitive qpan class="Chemical">uantum yield (44.4%). In addition, the reaction time and the purification process were dramatically shortened, avoiding the longer dialysis purification steps and sophisticated instrumentation (Table S1). Most importantly, the obtained GQDs provide new opportunities for the development of prominent applications due to their high comclass="Chemical">patibility with aqueous solutions, low cost, and class="Chemical">pan class="Disease">low toxicity. Furthermore, these nitrogen-containing GQDs open facile avenues for the modification with metallic cations and the development of competitive and selective fluorescent probes for the determination of pharmaceutical compounds.

Conflicts of interest

The authors declare that there are no conflicts of interest.